Risk Analysis of Collision Between Passenger Ferry and Chemical Tanker in the Western Zone of the Baltic Sea
Total Page：16
File Type：pdf, Size：1020Kb
POLISH MARITIME RESEARCH 2(86) 2015 Vol. 22; pp. 3-8 10.1515/pomr-2015-0011 RISK ANALYSIS OF COLLISION BETWEEN PASSENGER FERRY AND CHEMICAL TANKER IN THE WESTERN ZONE OF THE BALTIC SEA Marcin Przywarty, Ph. D. Lucjan Gucma, Assoc Prof. Krzysztof Marcjan, Ph. D. Andrzej Bąk, Assoc. Prof. Maritime University of Szczecin, Poland ABSTRACT This paper presents assumptions and process of the forming of a risk assessment model for collision between a passenger ferry departing from or approaching port of Świnoujście and a chemical tanker carrying a dangerous cargo. In order to assess navigational safety on the basis of data obtained from AIS system, were prepared probabilistic domains of ships, which made it possible to estimate number of navigational incidents as well as their spatial distribution , that consequently allowed to determine potentially dangerous areas. The next phase was formulation of a simulative model intended for the calculating of probability of collision between the ferry and chemical tanker as well as the determining of characteristic scenarios for such collision. This paper presents also an analysis of consequences of the collision with taking into consideration a damage of cargo tanks. Keywords: safety of navigation, navigational risk, ships domain, ships collisions, simulation methods Introduction In view of a large number of narrow passages and shallow waters the Baltic Sea has been always a difficult area for navigation. In spite of investment projects undertaken to improve navigational safety in the Baltic Sea any significant drop in number of navigational incidents has been observed so far. In order to obtain a comprehensive image of navigational safety over the Baltic Sea it is neccessary to have basic information on ship traffic flow. According to HELCOM data, total number of ships moving along certain routes per year exceeds 60 thousand. To improve navigational safety traffic separation systems are introduced. They organize ship traffic by lowering traffic flow breadth, that is associated with lowering the distance between passing-by ships. In certain areas the distances between passing-by ships drop below a value at which navigational incidents may occur. To estimate size and shape of ship domains for different kinds of encounters as well as to determine a number and spatial distribution of navigational incidents, in this work were examined 4 crossings of main Fig. 1. The investigated water area with depicted points where navigational navigational routes in the water area in question and routes of incidents occurred passenger ferries sailing between Świnoujście and Ystad (Fig.1). POLISH MARITIME RESEARCH, No 2/2015 3 Data concerning ship types and shipped cargoes, achieved The domain determined for encounter on crossing courses from AIS system, are limited to the code in which the first is of a shape most close to circle. It shows that both the ships digit stands for a type of ship (e.g. 6 – passenger ship, 8 – passing before bow of another ship (most often those having tanker), and the other digit – a type of shipped cargo: priority) and the ships passing behind stern of another ship • 1 –X category (most often those giving way) have maintained a comparable • 2 – Y category passing distance. The largest number for encounters on • 3 – Z category crossing courses occurred in the sector of 270° ÷ 315° • 4 – OS category (270 encounters, 20 incidents). The domain for overtaking is of an elongated shape both In this work occurrence probability was determined in fore and aft part. It results from the fact that for such of the most dangerous incident, i.e. collision between manoeuvre the points of the closest approach (CPA) occur a passenger ferry and chemical tanker carrying a dangerous usually when the overtaking ship is almost on the beam. cargo of X category, i.e. noxious liquids which, in case of their Most encounters (about 80%) identified as the overtaking, discharge to sea, could endanger sea resources or human reached CPA before reaching the beam position (270° ÷ 90°). health [2], and, consequences of such incident were also For encounters on opposite (head-on) courses majority of assessed. the closest approach points happened around a on- the-beam position, 80% out of which – when passing port side. However Preliminary assessment of navigational safety it is rather surprising that as much as 20% of ships passed each other starboard as the Regulation 14 of the International The first phase of this work has consisted in estimation of Regulations for Preventing Collisions at Sea [COLREG] states a number of navigational incidents on the basis of analysis that: if two ships go head to each other in such a way that it of actual data achieved from AIS system. To this end, for causes risk of collision then each of them is to so change its the selected water area probabilistic domains of ships were course as to be able to pass port side of the other. determined. Ship’s domain should be meant as a certain area On the basis of the determined domain parameters, around the ship whose navigator would like to keep free from number of navigational incidents which happened within other objects either permanent or moving ones. To calculate one month in the examined water area, was determined for parameters of the domains the methodology presented in the each type of the encounters. The incident (statistical one) publication [3] was used. It makes it possible to determine was defined as a situation in which ships pass each other limit distances in which majority of navigators (their number at a distance smaller than that determined for a given type is determined by 1st local maximum of distribution of CPA of encounter and a given course angle. Detail data on the distances) pass by other ships at different course angles in a incidents are presented in Tab. 1 given encounter situation. In Fig. 2 detail data dealing with shapes and sizes of the determined domains (marked blue), are Tab. 1 Incidents between ships in the examined water area presented. Additionally, the same domains, but diminished (marked green) and expanded (marked red) by 25%, are also depicted. In Fig. 1 positions of incidents are shown and, additionaly, a type of encounter at which a given incident has happened, is distinguished by a colour. Incidents on crossing courses occurred especially in the places where the route of ferries crosses navigational routes of a high traffic. Incidents during the overtaking occurred close to ports and in the places where routes cross each other. Incidents on opposite courses took place close to ports, that is associated with lowering the traffic flow breadth and, in consequence, lowering the distance between passing ships. The areas most dangerous for navigation, where the highest concentration of incidents Fig.2. Domains determined for particular types of encounters was observed, are determined by the following coordinates: 4 POLISH MARITIME RESEARCH, No 2/2015 • 55° 10.170’N - 55° 14.598’N and 13° 48.138’E - 13° 53.712’E • 54° 57.732’N - 55° 03.306’N and 13° 52.926’E - 13° 58.086’E • 54° 06.678’N - 54° 21.342’N and 14° 05.394’E - 14° 14.160’E Stochastic model of navigational safety To calculate occurrence probability of collision between passenger ferry and chemical tanker, a stochastic model of navigational safety assessment, developed by Institute of Sea Traffic Engineering, Maritime University of Szczecin, was implemented. Structure of the model is presented in Fig. 3. The model operates in the deliberately accellerated time flow in order to obtain a statistically sufficient number of scenarios. Fig. 4. The modelled navigational routes out at Ferry Terminal, Świnoujście [6]. The investigations consisted in determination of parameters of a theoretical distribution which allows to simulate delays of ferries or their earlier departures out of port. The following results were obtained from statistical analysis of the gathered data: Fig. 3. Schematic diagram of structure of the model for navigational safety assessment • The mean difference between actual and scheduled departure time of a ferry: 4min • Maximum difference between actual and scheduled departure time of a ferry: 34 min &KHPLFDOWDQNHUV¶WUDI¿FPRGHO • Minimum difference between actual and scheduled departure time of a ferry: -9min (the ferry got away from the Traffic of chemical tankers is modelled by means of pier prior to scheduled time) Poisson’s non-homogeneous distribution whose parameters • Modal value: 3min have been determined on the basis of actual data achieved • Standard deviation: 8,2 min from AIS system. Analysis of the data allowed to determine parameters of navigational routes as well as parameters of The detail data subjected to statistical analysis are shown distributions describing traffic flow along the routes (a.o. on the histogram (Fig. 5.) length, draught and speed). The so determined navigational routes are presented in Fig. 4. )HUULHV¶WUDI¿FPRGHO Traffic of ferries is modelled on the basis of an actual sailing schedule of ferries travelling on routes for and from Świnoujście. Ferries’ traffic parameters (a.o. speeds, turn points) as well as parameters of the ferries (a.o. values of their length, breadth and draught) are simulated on the basis of actual data taken from their technical documentation and also from AIS system. Differences in time of actual departures of ferries from a port in relation to scheduled ones are taken Fig. 5. Histogram of differences between actual and scheduled departure time into account by using a probabilistic model. The model has of ferries from Świnoujście been developed on the basis of actual observations carried POLISH MARITIME RESEARCH, No 2/2015 5 On the basis of characteristics of the random variable and calculated parameters as well as the developed histogram of the variable, for further analysis theoretical distributions capable of describing the given random variable, have been selected.